We used MLVA to genotype all Salmonella Typhimurium isolates referred to the NSW Enteric Reference Laboratory at the Centre for Infectious Diseases and Microbiology, NSW Health Pathology (Sydney, NSW, Australia), during August 2009–March 2016. We conducted multiplex PCR to amplify VNTRs (STTR9, STTR5, STTR6, STTR10pl and STTR3) and subsequent analyses as described previously (9,10). We reported MLVA results as a string of 5 numbers representing relevant repeats (11). We used Salmonella Typhimurium reference strain LT2 (GenBank accession on. NC_003197) as a control throughout and consistently generated the expected MLVA type 4-13-13-10-0211. We defined a cluster as >2 isolates with the same MLVA type collected within 12 weeks (10) and used a χ² test to determine the significance of differences observed; we considered a p value of <0.05 significant. We determined population diversity by calculating the Simpson index of diversity (12), and population richness using the McIntosh dominance index (13). This study was approved by our local Human Research Ethics Committee (LNR/17/WMED/25).

Figure 1

Figure 1. Trends of Salmonella enterica serovar Typhimurium notifications and multilocus variable-number tandem-repeat analysis (MLVA) patterns, New South Wales, Australia, 2009–2016. A) Quarterly counts of total cases and cases clustered by MLVA. B)...

After excluding duplicate isolates from the same episode of the disease (99% of all cases recorded in NSW), we examined a total of 15,626 human Salmonella Typhimurium isolates. We defined seasons as spring (September–November), summer (December–February), autumn (March–May), and winter (June–August). We observed a substantial fluctuation in the number of infections during the study period (Figure 1). The relative contributions of common definitive phage types (DT) and MLVA types to local Salmonella Typhimurium activity varied over 8 seasonal peaks. DT135 dominated in 2008 but was subsequently replaced by DT170 as the most common type in 2009 and 2010; an increase in the activity of DT9 occurred in 2014 and 2015. In total, we observed 667 different MLVA types. Six related STM DT170 MLVA types (2-7-6-12-0212; 2-7-7-12-0212; 2-7-6-13-0212; 2-7-6-11-0212; 2-7-6-14-0212 and 2-7-7-11-0212) represented 30% of all isolates. Increases in Salmonella Typhimurium cases during summer and autumn months (December–May in the Southern Hemisphere), with 2 cycles of 2009–2011 and 2012–2014, were mirrored by an increase in the number of unique MLVA types detected. The proportions isolates included in clusters also demonstrated expected seasonal fluctuations corresponding to increases in incidence (Figure 1, panel A), although we found no significant change in the number of clusters or their average size during the study period (p<0.05).

The average number of Salmonella Typhimurium cases for the 8 summer–autumn seasons investigated during the study period was 1,301. We categorized all summer–autumn seasons as either high (epidemic) or low on the basis of whether case numbers were above or below the average. We confirmed the designations by comparing the seasonal averages calculated for this study with yearly rates of salmonellosis notifications. These 2 approaches congruently assigned summer–autumn seasons of 2010, 2011, 2014, and 2015 as high or epidemic, whereas summer–autumn seasons of 2009, 2012, 2013, and 2016 were classified as low; the difference between case numbers in high and low seasons was significant (p<0.02) (Figure 1, panels B, C). The average annual number of foodborne community outbreaks recorded by the national public health network during high seasons was 63 and during low seasons was 48 (http://www.ozfoodnet.gov.au).

Figure 2

Figure 2. Population dynamics of Salmonella enterica serovar Typhimurium MLVA types, New South Wales, Australia, 2009–2016. A) Total number of novel or unique MLVA types. Red bars indicate high season and gray bars...

At the beginning of the study, we considered all MLVA types novel. The ratio of novel, previously unreported MLVA types to all types stabilized within 5 months; after the fifth month, 10%–40% of all MLVA types detected at any given time consisted of novel MLVA types (Figure 2). We detected no changes in the age distribution of human populations affected by dominant types; 41.1% of all infections occurred in those <14 years of age.

The diversity of the Salmonella Typhimurium population remained relatively constant over time; the McIntosh dominance index of diversity fluctuated between 0.6 and 0.9 during both high and low seasons (p = 0.478). However, we observed a rapid decrease in the proportions of unique MLVA types from winter to spring (i.e., U_w/U_sp ratio<1) before epidemic Salmonella Typhimurium activity. In contrast, the proportion of unique MLVA types increased from winter to spring preceding low seasons (U_w/U_sp ratio>1) (Figure 2). This ratio also correlated with incidence of Salmonella Typhimurium in NSW over the study period (r = 0.922). Of note, the percentage of unique MLVA types recovered from patients <14 years of age during winter of high/epidemic seasons was also significantly lower than during the winter months of low seasons of salmonellosis (15.1% vs. 28.2%; p<0.001).

Our observations potentially reflect the decrease in Salmonella Typhimurium diversity resulting from the reduction of genome variation under selection pressure, which is associated with the emergence of successful unique clones capable of causing epidemics in immunologically naive hosts. However, longer-term monitoring of subtype diversity and disease incidence is warranted to confirm these trends. Although an analogous selection-driven reduction in genetic diversity has been observed in other pathogens, such as influenza virus (14), it has not previously been observed in Salmonella. Our findings also suggest that the timely recognition of novel Salmonella Typhimurium subtypes may be of significance for surveillance and that the conventional diversity indices alone may not be sufficient to detect subtle changes in circulating subtypes. As the estimated ratio of accumulation of MLVA repeats in different loci to single nucleotide polymorphisms is 1:6.9 (15), changes in the composition of Salmonella Typhimurium subtypes might offer insight into the relevance of population diversity for fluctuations in the incidence of salmonellosis.

Substantial increases in seasonal epidemics of Salmonella Typhimurium can be associated with a reduction in newly identified MLVA types in the preceding winter and spring, reflecting the emergence of successful Salmonella Typhimurium clones under selection pressure. The proportion of novel MLVA types in winter and spring may serve as an early warning sign in public health surveillance. These observations add further insights into the epidemiology of Salmonella Typhimurium infections in a low-incidence setting. Although they may not be readily applicable to high-incidence Salmonella Typhimurium settings with frequent co-infections and different diagnostic or public health practices, the epidemiology of Salmonella Typhimurium and public health responses in Australia are similar to those in other industrialized countries, supporting the generalizability of our findings. Prospective monitoring of Salmonella Typhimurium population diversity and identifying new MLVA types as reservoirs from which future epidemics might emerge can improve the assessment of risks of seasonal increase in Salmonella Typhimurium incidence.